{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:YHS4S3AI2XMWSSZ7YL3HNUNGYU","short_pith_number":"pith:YHS4S3AI","schema_version":"1.0","canonical_sha256":"c1e5c96c08d5d9694b3fc2f676d1a6c52aa0c16daa83ff69995f50ea18247187","source":{"kind":"arxiv","id":"2111.14447","version":2},"attestation_state":"computed","paper":{"title":"ZeroCap: Zero-Shot Image-to-Text Generation for Visual-Semantic Arithmetic","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Idan Schwartz, Lior Wolf, Yoad Tewel, Yoav Shalev","submitted_at":"2021-11-29T11:01:49Z","abstract_excerpt":"Recent text-to-image matching models apply contrastive learning to large corpora of uncurated pairs of images and sentences. While such models can provide a powerful score for matching and subsequent zero-shot tasks, they are not capable of generating caption given an image. In this work, we repurpose such models to generate a descriptive text given an image at inference time, without any further training or tuning steps. This is done by combining the visual-semantic model with a large language model, benefiting from the knowledge in both web-scale models. The resulting captions are much less "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2111.14447","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-11-29T11:01:49Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"8497285d922229852e04128dae0786b43666c23f637cae1b9c9da45d7275c4b7","abstract_canon_sha256":"5d2e5f27558f797586229b0a48d9a11e3f50cb177c14d979e111fc4b64deb620"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:10:24.751580Z","signature_b64":"Y1hYF/sVlwBx2CHJmOHoaYRSJcfD2vyONMSM/8N55o5uY2q5UP/bVGN/QmzH3FQz9oRvGm4PLdahUzVDo5+tBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1e5c96c08d5d9694b3fc2f676d1a6c52aa0c16daa83ff69995f50ea18247187","last_reissued_at":"2026-07-05T04:10:24.751109Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:10:24.751109Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ZeroCap: Zero-Shot Image-to-Text Generation for Visual-Semantic Arithmetic","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Idan Schwartz, Lior Wolf, Yoad Tewel, Yoav Shalev","submitted_at":"2021-11-29T11:01:49Z","abstract_excerpt":"Recent text-to-image matching models apply contrastive learning to large corpora of uncurated pairs of images and sentences. While such models can provide a powerful score for matching and subsequent zero-shot tasks, they are not capable of generating caption given an image. In this work, we repurpose such models to generate a descriptive text given an image at inference time, without any further training or tuning steps. This is done by combining the visual-semantic model with a large language model, benefiting from the knowledge in both web-scale models. The resulting captions are much less "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.14447","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2111.14447/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2111.14447","created_at":"2026-07-05T04:10:24.751166+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.14447v2","created_at":"2026-07-05T04:10:24.751166+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.14447","created_at":"2026-07-05T04:10:24.751166+00:00"},{"alias_kind":"pith_short_12","alias_value":"YHS4S3AI2XMW","created_at":"2026-07-05T04:10:24.751166+00:00"},{"alias_kind":"pith_short_16","alias_value":"YHS4S3AI2XMWSSZ7","created_at":"2026-07-05T04:10:24.751166+00:00"},{"alias_kind":"pith_short_8","alias_value":"YHS4S3AI","created_at":"2026-07-05T04:10:24.751166+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.11074","citing_title":"Modeling Complex Behaviors: Multi-Personality Composition and Dynamic Switching in Vision-Language Models","ref_index":92,"is_internal_anchor":false},{"citing_arxiv_id":"2204.00598","citing_title":"Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language","ref_index":62,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU","json":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU.json","graph_json":"https://pith.science/api/pith-number/YHS4S3AI2XMWSSZ7YL3HNUNGYU/graph.json","events_json":"https://pith.science/api/pith-number/YHS4S3AI2XMWSSZ7YL3HNUNGYU/events.json","paper":"https://pith.science/paper/YHS4S3AI"},"agent_actions":{"view_html":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU","download_json":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU.json","view_paper":"https://pith.science/paper/YHS4S3AI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.14447&json=true","fetch_graph":"https://pith.science/api/pith-number/YHS4S3AI2XMWSSZ7YL3HNUNGYU/graph.json","fetch_events":"https://pith.science/api/pith-number/YHS4S3AI2XMWSSZ7YL3HNUNGYU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU/action/storage_attestation","attest_author":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU/action/author_attestation","sign_citation":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU/action/citation_signature","submit_replication":"https://pith.science/pith/YHS4S3AI2XMWSSZ7YL3HNUNGYU/action/replication_record"}},"created_at":"2026-07-05T04:10:24.751166+00:00","updated_at":"2026-07-05T04:10:24.751166+00:00"}